首页 | 本学科首页   官方微博 | 高级检索  
     检索      

大田经济作物性状研究中的图像分割技术应用
引用本文:陈 浩,任奕林,欧阳家乐,陈佃贞,王浩杰,徐 洋.大田经济作物性状研究中的图像分割技术应用[J].广东农业科学,2021,48(11):153-163.
作者姓名:陈 浩  任奕林  欧阳家乐  陈佃贞  王浩杰  徐 洋
作者单位:华中农业大学工学院/农业农村部长江中下游农业装备重点实验室,湖北 武汉 430070
基金项目:国家重点研发计划项目(2018YFD1000900);中央高校基本科研业务费专项资金(2662020GXPY004)
摘    要:大田经济作物在我国占有重要地位,其性状研究在机械化生产收获过程中至关重要,直接影响到生产效率和产品质量,数字图像技术的引入不仅为其性状研究带来高效的解决方案,更显示出广阔的应用前景。综合分析了近年来不断改进的图像分割方法,将其分成传统分割方法和结合特定工具的图像分割算法两大类,重点介绍了传统分割方法中基于阈值的分割方法和基于边缘检测的分割方法两种流行算法,结合特定理论工具的图像分割算法中的小波分析变换、遗传算法、主动轮廓模型、聚类算法、深度学习方法等 6 种方法,并对各大类算法中使用的各种图像分割方法进行归纳总结,列举其在油菜、棉花、大豆和花生等经济作物性状研究中的应用,同时对遗传算法、主动轮廓模型、聚类算法和深度学习中的 Fast R-CNN 算法添加技术框架图。最后对图像分割技术在大田经济作物中应用存在的主要问题进行分析并提出展望。

关 键 词:大田经济作物  图像分割  传统分割方法  结合特定工具方法  技术框架图

Application of Image Segmentation Technology in the Study of Field Cash Crop Characters
Abstract:Field cash crops occupy an important position in China, research on their characters is crucial in the process of mechanized production and harvest, which directly affect the production efficiency and product quality. The introduction of digital image technology not only brings efficient solutions to their characters research, but also shows broad application prospects. This paper analyzes the continuously improved image segmentation method, and divides it into traditional segmentation methods and image segmentation algorithm combined with specific tools. It mainly introduces the two popular methods, namely, segmentation method based on threshold and segmentation method based on edge detection, and wavelet analysis transform, genetic algorithm, active contour model, clustering algorithm and deep learning method combined with combined with specific tools. The image segmentation methods used in various algorithms are summarized, some applications in the characters of rape, cotton, soybean, and peanut are listed, and corresponding technicala framework diagram added in the genetic algorithm, active contour model, clustering algorithm and Fast R-CNN algorithm in deep learning. Finally, the main problems of the application of image segmentation technology in field cash crops are analyzed, and the prospects are put forward.
Keywords:field cash crop  image segmentation  traditional segmentation method  combined with specific tool methods  technical framework diagram
本文献已被 万方数据 等数据库收录!
点击此处可从《广东农业科学》浏览原始摘要信息
点击此处可从《广东农业科学》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号